Abstract #3350
Multiparametric Tumor Clustering for Predicting Recurrent Glioblastoma: Comparison with Single Parametric Diffusion and Perfusion Analyses
Ra Gyoung Yoon 1 , Ho Sung Kim 2 , Choong Gon Choi 2 , and Sang Jun Kim 2
1
Radiology, Asan medical center, Seoul,
Seoul, Korea,
2
Asan
medical center, Seoul, Korea
We performed this study to determine if enlarging
contrast-enhancing lesion (CEL) with similar tumor
microenvironment (TM) in patients with posttreatment
glioblastoma, can be labeled by clustering methods to
differentiate between recurrent glioblastoma (RGM) and
radiation necrosis (RN). The tumor clustering method
including four distinct clusters (tumor cluster,
radiation change cluster, necrosis cluster, edema
cluster) was performed on DSC, DCE, and DW MR imagings
of 84 patients with pathologically proven RGM or RN. We
have demonstrated that tumor clustering of clinical MR
imaging data is feasible. Moreover, the volume fraction
of tumor cluster was associated with the possibility of
RGM.
This abstract and the presentation materials are available to members only;
a login is required.
Join Here
